<Anchor>



Today (25th) starts with the news of plagiarism of a thesis by the artificial intelligence research team of Seoul National University.

It was submitted to a world-class conference and was also selected as an excellent paper, but it turned out that it was a plagiarism paper copied from several papers.

When allegations were raised, the authors immediately admitted their mistake and apologized.



First, reporter Jeong Seong-jin reports.



<Reporter Seongjin Jeong>



The CVPR Society is the world's most prestigious conference in the fields of artificial intelligence and machine learning.



It was held in the United States from the 19th to the 24th, and the thesis of the Seoul National University research team was selected as the excellent presentation thesis.



Enlarging an image


However, right after the announcement on the 23rd, a video accusing the thesis plagiarism was uploaded on YouTube.



I compared the Seoul National University thesis with several other papers, and in the 2019 University of Toronto thesis, Canada, the same sentence was found without changing a single word, and the formula, which is the core of the thesis, has the same structure as the thesis of an overseas research group published last year.



Enlarging an image


There are absolutely no citations to these papers.



There are close to 10 copies of theses, including a thesis from the University of Berkeley in the US in 2018, a thesis from the University of Toronto in Canada in 2019, a KAIST thesis in 2021, and a thesis from Oxford in the UK.



One day after the allegations were raised, four people, including the first author and co-author of the paper, officially admitted plagiarism.



The first author, Mo Kim, said that he was entirely at fault, and that he would accept any discipline.



The three co-authors said, "The first author did the main writing of the thesis. However, we feel responsible and apologize for not finding plagiarism."



The first author, Researcher Mo Kim, who attended the CVPR conference alone, is planning to return to Korea soon to participate in the Seoul National University investigation.



(Video coverage: Seol Min-hwan, video editing: Choi Hye-young)



---



<Anchor>



Plagiarism The corresponding author is Professor Yun Seong-ro of Seoul National University.

He served as Chairman of the 4th Industrial Revolution directly under the President and was recently selected as a 'Leader Researcher' receiving government research grants.

That is why he is regarded as an authority in academia.

He asked Professor Yoon why he didn't notice the plagiarism.



This is reporter Jeong Da-eun.



<Reporter Jeong Da-eun>



The name of the corresponding author of the plagiarism thesis is Professor Yun Seong-ro of Seoul National University's Artificial Intelligence Research Center.



Enlarging an image


Professor Yoon served as the Chairman of the 4th Industrial Revolution under the direct control of the President during the Moon Jae-in administration.



After the inauguration of the current government, he was selected as a 'leader researcher' by the Ministry of Science and Technology, which provides annual research funding of around 800 million won.



Professor Yoon, who remained in Korea, said in a phone call with SBS, "After confirming the fact of plagiarism, I requested the CVPR Society to withdraw the thesis. .



The CVPR side also posted an article on SNS saying "plagiarism is unacceptable" and has requested an investigation from the World Institute of Electrical and Electronic Engineers (IEEE).



Regarding the reason for not checking plagiarism in advance even though the corresponding author was the corresponding author, Professor Yoon explained, "The corresponding author and co-author review the contents of the paper with the first author, but they do not check whether plagiarism is detected by comparing them one by one with other papers." I did.



However, criticism from domestic researchers continues.



On the SNS, evaluations such as "a garbage-like behavior that pours cold water on the efforts of other researchers" and "an incident that occurred when a domestic machine learning paper was made in the form of printing" were also raised.



With this incident, the status of CVPR, the most prestigious academic society, is shaken.



Overseas scholars are pointing out that the process of selecting Korean plagiarized papers as excellent publications without sufficient verification should be reviewed again.



(Video coverage: Kim Se-kyung, video editing: Yu Mira, CG: Park Cheon-woong)

Keywords: